Download
VideOCR extracts hardcoded subtitles from video files using machine learning and optical character recognition. It provides a graphical interface for locating subtitle regions and converting visible text into editable subtitle files. Users can navigate video timelines and select a specific extraction area using a cropping rectangle. The application supports more than 200 languages through its OCR capabilities. Local processing uses PaddleOCR, while an optional Google Lens mode provides cloud-assisted recognition. Advanced settings control extraction behavior, including frame sampling and processing parameters. GPU acceleration can improve performance on compatible hardware. A command-line version supports automated processing and Docker deployment. VideOCR runs on Windows and Linux.
Features
- Automatic extraction of burned-in video subtitles using optical character recognition
- Multilingual subtitle recognition supporting more than 200 languages
- Offline PaddleOCR processing with optional Google Lens cloud-assisted recognition
- Interactive video timeline navigation with customizable subtitle detection regions
- GPU-accelerated subtitle extraction with adjustable frame sampling and timing settings
- SRT subtitle exporting through graphical, command-line, and Docker-based workflows
Project Samples
